Coventry University Repository for the Virtual Environment (CURVE)

Transcription

1 Coventry University Coventry University Repository for the Virtual Environment (CURVE) Author names: Pasiouras, F., Tanna, S. and Gaganis, C. Title: What drives acquisitions in the EU banking industry? The role of bank regulation and supervision framework, bank specific and market specific factors. Article & version: Pre-print version Original citation: Pasiouras, F., Tanna, S. and Gaganis, C. (2011) What drives acquisitions in the EU banking industry? The role of bank regulation and supervision framework, bank specific and market specific factors. Financial Markets, Institutions & Instruments, volume 20 (2): Publisher statement: This is the pre-peer reviewed version of the following article: Pasiouras, F., Tanna, S. and Gaganis, C. (2011) What drives acquisitions in the EU banking industry? The role of bank regulation and supervision framework, bank specific and market specific factors. Financial Markets, Institutions & Instruments, volume 20 (2): 29-77, which has been published in final form at Copyright and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

2 This document is the submitted version of the journal article, as originally submitted to the journal prior to the peer-review process. There may be some differences between the published version and this version and you are advised to consult the published version if you wish to cite from it. Available in the CURVE Research Collection: August

3 What drives acquisitions in the EU banking industry? The role of bank regulation and supervision framework, bank specific and market specific factors Fotios Pasiouras 1,2*, Sailesh Tanna 3, Chrysovalantis Gaganis 4 1 Deparment of Production Engineering & Management, Technical University of Crete, Greece 2 Centre for Governance & Regulation, University of Bath, UK 3 Department of Economics, Finance & Accounting, Faculty of Business, Environment & Society, Coventry University, UK 4 Department of Economics, Faculty of Social Sciences, University of Crete, Greece Abstract We investigate the determinants of commercial bank acquisitions in the former fifteen countries of the European Union by evaluating the impact of bank-specific measures, such as size, growth and efficiency of banks, and external influences reflecting industry level differences in the regulatory and supervision framework, market environment and economic conditions. Our empirical analysis involves multinomial logit estimation at various levels in order to identify those characteristics that most consistently predict targets and acquirers from a sample of over 1400 commercial banks. The overall results indicate that, relative to banks that were not involved in the acquisitions, (i) targets and acquirers were significantly larger, less well capitalized and less cost efficient, (ii) targets were less profitable with lower growth prospects, and acquirers more profitable with higher growth prospects, (iii) external factors have affected targets and acquirers differently, and their effects have not been consistent or robust to sample size changes. Keywords: Acquisitions, Banks, Logistic regression, Regulations, Supervision JEL: G21, G28, G34 * Corresponding author. Tel/Fax: /69410; s: pasiouras@dpem.tuc.gr, f.pasiouras@bath.ac.uk (F. Pasiouras) s.tanna@coventry.ac.uk (S. Tanna); c.gaganis@econ.soc.uoc.gr (C. Gaganis). We would like to thank participants at the 5 th INFINITI Conference on International Finance (Dublin, 2007), and the 2010 EWG-EPA International Conference (Chania, 2010) for valuable comments that helped us to improve earlier versions of the manuscript. 1

4 I. INTRODUCTION The EU banking industry has witnessed a large number of mergers and acquisitions (M&As) in recent years. The European Central Bank (2000), for example, records 2,153 M&As of credit institutions between 1995 and the first half of 2000, while Beitel and Schiereck (2001) point out that during the period more M&As deals occurred in the EU banking industry than during the previous 14 years. In terms of volume, data from the Securities Data Company (SDC) M&A Database 1 indicate that the total value of European financial M&As increased from $22,769.6 million in 1990 to $147,025.6 in 1999, while over the same period, the average target value in Europe ($467.7 millions) was higher than in the US ($334 millions) and the main industrial countries on an aggregate basis ($383.2 millions). Theory suggests that M&As between banks can occur for several reasons. In general, the underlying motives can be classified as value-maximization (i.e. increase market power, replace inefficient management, achieve economies of scale and scope, decrease risk through geographic and product diversification) and non-value maximization ones (i.e. managerial motives, hubris, etc.). In addition to these firm level motives, banks decision for M&As might be influenced by external factors such as industry level differences in the economic environment, laws, regulations, etc (Berger et al., 1999; Group of Ten, 2001). While there are numerous empirical studies investigating the relationship between financial characteristics and acquisition likelihood of industrial (i.e. non-financial) firms (e.g. Levine and Aaronovitch, 1981; Harris et al., 1982; Hasbrouck, 1985; Ambrose and 1 Produced by Thomson Financial Securities Services. 2

5 Megginson, 1992; Powell, 1997; Gonzalez et al., 1997; Ali-Yrkko et al., 2005), investigation of such characteristics for the banking industry has been limited (Cyree et al., 2000; Wheelock and Wilson, 2000). Furthermore, previous studies on bank acquisitions have traditionally focused on examining the financial characteristics of US banks (Hannan and Rhoades, 1987; Meric et al., 1991; Moore, 1996; Wheelock and Wilson, 2000, 2004; Hannan and Pilloff, 2009), while there have been relatively few studies for the EU countries. Hernando et al (2009) and Kohler (2009) are the most recent focussing on domestic and cross-border bank acquisitions in the EU-25 countries, while Lanine and Vander Vennet (2007) examine the characteristics of cross-border acquisitions of Central and Eastern European banks by Western European banks. 2 In addition, a limited number of studies have focussed on specific countries of the EU (e.g. Focarelli et al, 2002; Pasiouras and Zopounidis, 2008; Pasiouras and Gaganis, 2009). Evidence on the impact of external factors on M&As decisions also comes mostly from studies that examine industrial sectors, with the neoclassical and behavioural approaches being the most commonly cited explanations. From the previously mentioned studies in the banking sector, some have examined the impact of the economic environment using industry level characteristics such as market concentration, growth, profitability or size. However, the neoclassical theory, proposed by Gort (1969) and more recently supported by Mitchell and Mulherin (1996) among others, assumes that legal and regulatory factors might also have a role to play in the reallocation of corporate assets through M&A activity. Thus, for example, Rossi and Volpin (2004) examine the 2 These three studies for the EU, as well as Hannan and Pilloff (2009) for the US, came to our light since we wrote the initial version of this paper. Hernando et al (2009) actually refer to evidence reported in this paper but, as they acknowledge, our paper differs from theirs by providing evidence for the EU-15 with a focus on the role of differences in regulations. 3

6 influence of differences in law and regulation in their study of the determinants of M&As across 49 major countries, and find that the volume of M&A activity is significantly larger in countries with better accounting standards and stronger shareholder protection. In the bank M&As literature, the study of the impact of regulations and supervision approaches has also been investigated as forces hindering cross-border deals. Focarelli and Pozzolo (2001), using data on 2,449 banks from 29 OECD countries, point out that cross-border M&As among banks are less frequent than in other sectors of the economy, and find that the difference depends partly on the level of regulatory restrictions. In another study, Focarelli and Pozzolo (2005) examine where banks expand their cross-border shareholdings and find that potential profit opportunities and regulatory environments are the most important determinants. Buch and DeLong (2004a) provide further evidence on why cross-border mergers are rare compared to domestic mergers using a large sample of over 3000 international bank M&As. Treating the number of cross-border bank mergers for each country pair as the dependent variable in Tobit regressions, they find that information costs and regulations significantly influence crossborder merger activity. In their later study of cross-border bank mergers for the OECD countries, Buch and DeLong (2004b) reveal that a fairly priced deposit insurance scheme in the acquirer s country tends not only to increase the number of cross-border deals but also reduce the risk in both the home and world markets. More recently, Kohler (2009) examines the impact of merger control as a potential deterrent to EU bank acquisitions and finds that the transparency of the merger approval process serves to influence the likelihood of cross-border acquisitions, although domestic deals are unaffected by this. 4

7 Studies for the US banking industry are by their very nature limited to domestic M&As, although Wheelock and Wilson (2004) examine the impact of state branching laws and regulator evaluations of banks safety and soundness, focusing principally on acquirers. Their results also indicate that the regulatory approval process serves as a real constraint on bank merger activity, although changes in branching restriction are not statistically significant. This paper adds to the recent literature by investigating the acquisition likelihood characteristics for the EU banking industry. As noted above, relative to the US, the literature investigating the characteristics of bank acquisition likelihood in the EU has been limited, and we attempt to provide further evidence by concentrating on the period , when M&A activity in the EU banking industry was intense. 3 Our dataset consists of industry level data on the first 15 EU countries (EU15), and financial data for over 1,400 commercial banks operating in EU15, these being distinguished as acquirers, targets and non-involved banks. This unique dataset therefore enables us to analyze the ex-ante characteristics of both acquired and acquiring banks relative to non-acquired peers. 4 In doing so, we concentrate on evaluating the relative influence of bank level characteristics and industry level differences in the banks operating environment and economic conditions, as well as in their regulatory and supervision frameworks. 3 We concentrate on this period because it witnessed a reduction of nearly 23% in the number of banks in the EU and this decrease was due largely to domestic bank M&As as banking groups consolidated their position within countries to create national champions (ECB, 2004; Campa and Hernando, 2006). Hence, our sample includes mainly domestic deals, as cross-border integration in banking remained limited until recently (ECB, 2008). See Hernando et al (2009), Lanine and Vander Vennett (2007) and Kohler (2008) for recent evidence relating to cross-border bank M&As within the EU. 4 In this sense, we analyse both the pull and push factors affecting the probability of acquisition. Note that the terms acquirer (or bidder), target and non-involved could alternatively be interpreted as acquiring, acquired and non-acquired banks respectively, and will be used interchangeably in this paper. 5

8 The distinguishing aspect of our study is the examination of a broad range of policy influences that proxy for bank regulations and supervision standards, such as the level of accounting and information disclosure requirements, the degree of official disciplinary power, deposit insurance schemes, capital adequacy requirements, restrictions on bank activities and diversification guidelines. We obtain this information from the World Bank database, developed by Barth et al. (2001), and presume that these policy variables have either a direct impact on M&As or an indirect impact, for example, by limiting the investment opportunities of banks or influencing their risk-taking behaviour. 5 As noted above, Focarelli and Pozzolo (2001, 2005) and Buch and De Long (2004a,b) consider the impact of the regulatory environment on cross border deals. In contrast, we consider the impact of country-specific differences in the regulatory environment on commercial bank M&As in the EU single market, where such deals have been largely domestic. Although this makes our study somewhat related to the studies of Rossi and Volpin (2004) and Wheelock and Wilson (2004), it should be noted that the former uses, in the main, the volume of merger activity as a dependent variable 6 and does not focus on the banking industry, while the latter concentrates on investigating characteristics of US bank acquirers that originate from the CAMEL approach with a limited set of further attributes to represent market environment and regulations. 5 Many studies argue that regulations such as capital requirements, deposit insurance scheme, restrictions on bank activities, disciplinary power of the authorities can have an impact on the risk taking behavior of banks (e.g. Besanko and Kanatas, 1996; Demirguc-Kunt and Kane, 2002; Hovakimian et al, 2003; Fernandez and Gonzalez, 2005; Gonzalez, 2005; Pennacchi, 2006). Amihud et al. (2002) and Buch and DeLong (2004b) point out that one way to take advantage of such regulations is to acquire a risky bank. 6 With the exception of recent studies (e.g. Lanine and Vander Vennett 2007; Hernando et al, 2009; Koehler, 2008), most previous studies that examined cross-border mergers focussed on the number of mergers (i.e. activity) rather than on the probability of individual banks to engage in M&As. 6

9 Using multinomial logit estimation to determine the impact of the above factors on the probability of acquisition, we show, with a fair degree of consistency across various levels of estimation, that both targets and acquirers were significantly larger, less well capitalized and less efficient in terms of expenses management, relative to their nonacquired peers. Furthermore, targets were less profitable with lower liquidity and lower growth in total assets; whereas acquirers tended to be relatively more profitable banks with higher growth prospects. These bank-specific influences are invariant to robustness tests conducted by disaggregating the sample according to bank size, location of operation and different time periods. But the impact of the regulatory and market environments are not robust to these sample splits and therefore depends crucially on whether the banks involved in acquisitions were large or small, and specifically where they operated. Besides, some regulatory influences were not uniform on targets and acquirers. Nevertheless, we find supporting evidence to suggest that banks that operated in countries with higher disciplining power of the authorities were less likely to engage in acquisitions, as targets or acquirers. Similarly, banks were more inclined to engage in acquisitions in market environments favouring higher profitability, higher liquidity, lower concentration and lower industry size, although these influences were not robustly significant. Furthermore, regulatory factors were found to have a greater influence on banks acquisitions in the principal banking sectors (i.e. the five largest countries of the EU) than in the rest of the EU-15 where market influences were more prevalent. The rest of this paper is organized as follows. Section II presents a review of prior literature related to our study. Section III outlines the data and methodology, while 7

10 Section IV discusses the empirical results. Finally, Section V outlines some concluding remarks. II. BACKGROUND DISCUSSION In this section we provide a comprehensive review of the relevant literature in order to justify the importance of using appropriate controls for bank regulation and supervision standards, and market-related economic conditions associated with M&As decisions in the banking industry, in addition to relevant bank specific characteristics. The discussion is split into three sub-sections, referring to each of the three broad categories in turn. Bank M&As and bank specific characteristics The causes of M&As have long been debated in the literature. Following the neoclassical perspective, all firm decisions including acquisitions are made with the objective of maximizing shareholders wealth. M&As in this context serve as a means to increase market power, replace inefficient management, achieve economies of scale and scope, decrease risk through geographic and product diversification, among others. However, an influential view in the literature is that M&As are driven by agency conflicts of interest between managers and shareholders. According to this view, many acquisitions are undertaken by managers in order to enhance their salary and prestige, diversify personal risk or secure their job through empire-building, at the expense of shareholders. Another interesting hypothesis, proposed by Roll (1986), suggests that managers commit errors of over-optimism (hubris) in evaluating M&As opportunities due to excessive prediction or 8

11 faith on their own abilities, and engage in M&As even when there is no synergy. The empirical evidence is inconclusive and indicates that various bank specific and market environmental factors can influence M&As in the banking industry. Hence, while the discussion below is devoted to financial and environmental characteristics, non-financial attributes relating to managerial incentives and ownership control may be important too (Hadlock et al., 1999; Brook et al., 2000; Bliss and Rosen, 2001; Hughes et al., 2003). 7 Capital Strength Harper (2000) argues that The key factor driving mergers and acquisitions in financial systems is the industry s need to rationalize its use of capital (p. 68). This argument is based on the belief that nowadays risks are traded on markets rather than absorbed through capital held on a balance sheet. Hence, in order to remain competitive banks face the need either to release surplus capital or to raise the rate of return to the capital they retain. This can be achieved through M&As. Most of the studies report a negative relationship between capital ratios and the likelihood of being acquired although not statistically significant in all cases (e.g. Hannan and Rhoades, 1987; Moore, 1996; Wheelock and Wilson, 2000; Lanine and Vander Vennet, 2007). There are several explanations for this. First, lack of financial strength tends to attract well capitalised buyers that can infuse capital into the acquired banks. Specifically, banks that are generally close to failure are encouraged by the authorities to 7 Lack of appropriate data precludes investigation of these issues given our comprehensive sample of public and private banks. However, agency cost influences have been linked to industries where hostile or diversifying acquisitions are more prevalent, as diversification benefits managers due to the diversification discount (Morck et al, 1990), but these types of acquisitions are rare in banking owing to regulatory hurdles. 9

12 be taken over by well capitalized banks. Second, better capitalised banks would be less attractive to potential buyers if capitalization is seen to indicate managerial efficiency. Third, buyers are attracted by less well capitalised banks with skilful managers who show ability to operate successfully with high leverage. Related to the third argument, Hannan and Piloff (2009) suggest that buyers prefer poor capitalized targets because it enables them to maximize the magnitude of post-merger performance gains relative to the cost of achieving these gains. Banks may also engage in M&As to meet higher capital regulatory requirements, suggesting a positive link between capitalization and acquisition likelihood. Valkanov and Kleimeir (2007) examine a sample of US and European bank mergers and find that US targets are better capitalized than acquirers and non-acquired peers and that US banks maintain higher capital levels than European banks. They suggest that US banks strategically raised their capital levels through mergers to avoid regulatory scrutiny. Alternatively, as suggested by Hernando et al. (2009), if capitalization signifies the inability of a bank to diversify assets, more capitalized banks would be worth more to better diversified acquirers, thus enhancing the likelihood of being acquired. Hernando et al. (2009) discuss these hypotheses about positive and negative links between acquisition and capitalization but find the effect of the latter insignificant in both domestic and crossborder samples. Performance According to the inefficient management hypothesis, acquisitions serve to drive out bad management that is not working in shareholder interests. Thus, as discussed by Hannan and Rhoades (1987), poorly managed banks are likely targets for acquirers who think that 10

13 they can manage more efficiently the assets of the acquired bank and increase profits and value. This outcome is more likely in domestic (or in-market) than in cross-border (outof market) M&As, because a local acquirer may be in a better position to turn around the fortunes of the target bank (Hernando et al, 2009). However, the empirical results are mixed. Moore (1996), Focarelli et al. (2002), Wheelock and Wilson (2000), Pasiouras and Gaganis (2009), Hannan and Pilloff (2009) and Hernando et al (2009) find evidence of a negative association between target performance (measured in terms of either return on assets (profitability), expense ratios such as cost-to income, or both) and acquisition likelihood. In contrast, Hannan and Rhoades (1987), Pasiouras and Zopounidis (2008) and Lanine and Vander Vennet (2007) find no evidence of such association, whereas Kohler (2009) finds the effect of targets return on assets significant in cross-border deals only, indicating that profitability does not seem to influence the probability of being acquired in domestic M&As. 8 Size Size may influence M&As in several ways. First, large banks are more expensive to be acquired. Second, larger banks have greater recourse to fight hostile acquisitions, as well as resources to acquire other banks. Third, a larger acquired bank is likely to be more difficult to be absorbed in the existing organization of the acquiring bank. These considerations suggest that the coefficient of size (as measured by total assets) on acquisitions should be negative. On the other hand, an acquirer seeking economies of scale or market power may find larger bank targets more attractive. 8 However, using the cost-to-income ratio instead of the return on assets, Kohler (2008) claims to find the effect significant in both domestic and cross-border cases. 11

14 Hannan and Rhoades (1987) and Moore (1996) find the effect of size insignificant. Wheelock and Wilson (2000), however, report that smaller banks are more likely to be acquired than larger ones, while Wheelock and Wilson (2004) find that the acquirers probability of engaging in mergers increases with bank size. Focarelli et al. (2002) report a negative and statistically significant effect of size on acquisitions for Italy, while Pasiouras and Zopounidis (2008) find a negative, though not robustly significant, effect for Greece (depending on the measure of size). In contrast, Hannan and Pilloff (2009) report a positive impact of size in their full sample, but a negative effect on a sub-sample focusing on smaller acquirers. Lanine and Vander Vennet (2007) and Kohler (2009) report a positive and significant impact of size in all their samples, although Hernando et al (2009) find a similar result only for domestic deals. Growth Bank growth can affect bank acquisition in two opposing ways. On the one hand, as Kocagil et al. (2002) point out, empirical evidence suggests that some banks with relatively high growth rates experience problems because their management and/or structure is not able to deal with and sustain exceptional growth. Hence, acquirers may purchase a bank with good growth prospects, but with limited financial or managerial capacity may fail to capitalize on potential growth. On the other hand, Moore (1996) argues that a slow growing bank may attract a buyer seeking to accelerate its growth rate and thereby increase its market value. Hannan and Rhoades (1987) find growth to be positively related to in-market acquisitions and negatively related to out-of-market acquisitions, albeit insignificant in 12

15 both cases. However, Moore (1996) and Pasiouras and Zopounidis (2008) find asset growth to be negatively related to acquisition likelihood. Pasiouras and Gaganis (2009) also find growth to be negatively related to the acquisition likelihood but statistically significant only in the case of Germany and Spain. Wheelock and Wilson (2004) report that acquirers tended to have a recent history of rapid growth. Among the more recent studies, Hannan and Piloff (2009) and Koehler (2009) do not include asset growth in their regressions, while Hernando et al (2009) find its effect insignificant on acquisitions. Loan activity The importance of loans for EU banks becomes apparent when reviewing data from the European Central Bank (2004) on the stability of the EU banking sector, which indicates that the share of customers loans in total assets was 50.57% in Therefore loan activity may be another factor influencing the decision to acquire a bank. Hannan and Rhoades (1987) argue that, on the one hand, a high level of loans would seem to indicate aggressive behaviour by the target bank and a strong market penetration with important established customer relationships that would make it an attractive target; whereas, on the other hand, a low level of loan activity may indicate a bank with conservative or complacent management, which an aggressive acquiring bank could turn around to increase returns. Hannan and Rhoades (1987) find a negative effect of loan activity on acquisition likelihood (although not significant). Moore (1996) also finds a negative (and significant) effect in both in-market and out-of-market acquisitions. The results of Wheelock and Wilson (2000, 2004) are generally mixed depending on the measure of asset quality used. 13

16 Pasiouras and Zopounidis (2008) also find this effect to be negatively related to the probability of acquisition, although not statistically significant in all cases, while Pasiouras and Gaganis (2009) report mixed results across the five large EU countries. Finally, none of the recent studies include a measure of loan activity, although Hannan and Piloff (2009) consider two measures of the composition of the target s clientele, proxied by the extent to which their loans (and deposits) are local in nature, and find the effect of the loans ratio to be positive and significant only in the case of large acquirers, while that of deposits positive and significant in all but one specification. Liquidity The liquidity position of a bank is another factor that may influence its attractiveness as an acquisition target. However, it is difficult to determine a priori what the effect of liquidity and the direction of its influence will be. Without the necessary liquidity and funding to meet obligations, a bank may fail unless external support is given (Golin, 2001). Hence, banks might be acquired because they have moved into liquidity difficulties, indicating that low liquidity increases acquisition likelihood. On the other hand, excess liquidity may signal a lack of investment opportunities or a poor allocation of assets, making banks attractive targets because of their good liquidity position (i.e. the size of liquid assets influences acquisition). Among the studies that include this variable, Wheelock and Wilson (2000) find that low liquidity makes banks less attractive targets, thus providing support to the first argument, while Pasiouras and Zopounidis (2008) report a negative relationship between liquidity and acquisition likelihood although not statistically significant. 14

17 Bank M&As and regulations and supervision Capital requirements In their assessment of the likely impact of Basel II capital requirements on bank M&As, Hannan and Pilloff (2004) propose that regulatory capital can affect acquisition activity in one of two ways. Under the excess regulatory capital hypothesis, merger activity would increase as a result of the excess regulatory capital that would be created by the lower capital requirements stemming from the adoption of advanced internal ratingsbased (A-IRB) approach to regulatory capital requirements 9. Under the relative capital advantage hypothesis, as a result of differences in the capital standards applied to A-IRB banks and other banking organizations not using the A-IRB approach, A-IRB banks would acquire banks not subject to A-IRB standards because acquired banks would be worth more to A-IRB banks than to current owners. They use data from US banks to test the excess regulatory capital hypothesis but do not find convincing evidence to suggest that past changes in excess regulatory capital or past changes in capital standards had substantial effects on merger activity. However, Valkanov and Kleimeier (2007) find evidence to support the excess regulatory capital hypothesis. Following an event study methodology, they find that more value is created for targets with high excess capital and in M&As involving targets with considerably higher excess-capital ratios than their acquirers. 9 This can occur for two reasons. First, while regulators may prevent banks with no excess regulatory capital to engage in M&As as the combined entity might violate minimum capital adequacy standards, banks with levels of regulatory capital above the required minimum are less likely to violate minimum standards, increasing the probability to acquire other banks. Second, with an increase in excess regulatory capital, banks should increase their return on equity either by increasing the amount of earning assets against which a given amount of capital is held or by reducing capital held against a given amount of earning assets. This could result in an increase of banks valuation, leading to an increase in acquisition activity. 15

18 Capital requirements can also have indirect effects on M&As through their impact on the risk-taking behaviour of banks. The main argument in support of capital requirements is that capital serves as the last line of defence against the risk of bank s insolvency, as any losses a bank suffers could be potentially written off against capital. Even in the case where insolvency becomes unavoidable, capital protects to some degree depositors, creditors and investors (Le Bras and Andrews, 2004). However, another strand of the literature indicates that capital requirements may increase risk-taking behaviour (e.g. Koehn and Santomero, 1980; Besanko and Kanatas, 1996; Blum, 1999; Calem and Rob, 1999). Other studies provide mixed results. Kendall (1992) suggests that higher capital requirements may cause riskier bank behaviour at some points in time, but do not imply a trend towards a riskier banking system. Beatty and Gron (2001) indicate that capital regulatory variables have significant effects for low-capital banks but not necessarily for other banks. Restrictions on bank activities Barth et al. (2004) outline several theoretical reasons for restricting bank activities as well as alternative reasons for allowing banks to participate in a broad range of activities. For example, they mention that to the extent that moral hazard encourages riskier behaviour, banks will have opportunities to increase risk if allowed to engage in a broader range of activities (Boyd et al., 1998). Furthermore, large financial conglomerates may reduce competition and efficiency. On the other hand, fewer regulatory restrictions permit the utilization of economies of scale and scope (Claessens and Klingebiel, 2000), while they 16

19 might also increase the franchise value of banks and result in more prudent behaviour. Finally, broad activities may enable banks to diversify income streams. Hence, higher restrictions on bank activities that will affect banks opportunities to diversify risks, and limit the potential for economies of scope and scale, might influence their investment decision by motivating them to engage in M&As as an alternative way to achieve their desired outcomes. Diversification and liquidity related regulations As Liang and Rhoades (1991) mention, a predicted benefit of mergers, particularly conglomerate mergers, is that diversification across different markets will reduce a firm s risk. For example, Liang and Rhoades (1988) point out that geographic diversification potentially permits banks to reduce their insolvency risk primarily through reduction in credit and liquidity risk. However, banks might achieve diversification by following alternative strategies such as making loans abroad or investing in various liquid assets. Hence, regulations that encourage or restrict banks with respect to liquidity as well as asset geographical diversification might also have an impact on M&As. Deposit insurance The literature suggests that the deposit insurance scheme of a country can have an impact on the risk behaviour of banks and their investment decisions. For example, Krugman (1998) suggests that banks that are over-guaranteed and under-regulated tend to overinvest. Other studies indicate that deposit insurance schemes may encourage excessive risk-taking behaviour (Merton, 1977; Bhattacharya and Thakor, 1993; Bhattacharya et al., 17

20 1998; Hendrickson and Nichols, 2001; Demirguc-Kunt and Kane, 2002). The deposit insurance scheme might also have an effect on the stability of the banking systems as a whole (Demirguc-Kunt and Detragiache, 2002; Barth et al., 2004). However, Kane (2000), Cull et al. (2005), Demirguc-Kunt and Detragiache (2002), Demirguc-Kunt and Kane (2002), and Laeven (2002) conclude that a sound legal system with proper enforcement of rules reduces the adverse effects of deposit insurance on bank risk-taking, while Gonzalez (2005) finds that deposit insurance has a positive influence on bank charter value, mitigating the risk-shifting incentives it creates. Finally, as noted earlier, Buch and Delong (2004b) find that fairly priced deposit insurance in the acquirer s country tends to increase the number of mergers banks participate in. Disclosure requirements Rossi and Volpin (2004) argue that accounting and information disclosure requirements may affect M&As because good disclosure is a necessary condition for identifying potential targets. They also argue that accounting standards reveal corporate governance as they decrease the scope for expropriation by making corporate accounts more transparent. Their empirical results indicate that the volume of M&A activity is significantly larger in countries with better accounting standards, hence providing support to their argument. Disclosure requirements may also have an impact on the risk-taking behaviour of banks, and consequently on their investment behaviour (e.g. M&As). For example, Fernandez and Gonzalez (2005), Demirguc-Kunt et al. (2008), and Agoraki et al. (2010) 18

21 find evidence that accounting and auditing requirements can improve the soundness of banks and decrease their risk-taking. Disciplinary power of supervisory agency Buch and DeLong (2004b) point out that weak supervision could alter banks decision making by fascinating them to engage in risky activities while ignoring activities that make good business sense. Obviously, one way to take advantage of a weak supervision system is to acquire a risky bank. Looking at cross-border M&As, Buch and Delong (2004a) point out that a tough supervisory system in the target country increases the number of bank mergers, while greater toughness of the acquiring country s authorities discourage mergers. The disciplinary power of supervisor agencies might also have an indirect impact on M&As, through its influence on banks performance and development. While the results of Barth et al. (2004) indicate that there is not a strong association between bank development and performance and official supervisory power, Fernandez and Gonzalez (2005) report that in countries with low accounting and auditing requirements a more stringent disciplinary capacity of supervisors over management action appears to be useful in risk reduction. Overall country s legal environment and openness Banks will obviously be affected by the overall environment of the country in which they operate, with a number of aspects relating to the environment having an impact on their investment decisions. For example, La Porta et al. (1998) and Levine (1998) among 19

22 others mention the effects of differences in the legal environments on the financial system. Rossi and Volpin (2004) show evidence of more M&A activity in countries with better investor protection. Furthermore, Francis et al. (2008) find some evidence that the abnormal returns of U.S. acquirers involved in cross-border M&As depend on economic freedom. Other studies show that the influence of regulations on banking is conditional on the political and economic environment. For example, Hovakimian et al. (2003) find that the introduction of explicit deposit insurance has adverse effects in environments that are low in political and economic freedom and high in corruption. Similarly, Barth et al. (2004) show that better-developed private property rights and greater political openness mitigate the negative association of moral hazard and bank fragility. Finally, Fernandez and Gonzalez (2005) also report that banks in a poor legal system with improper enforcement of rules carry a higher risk. Bank M&As and market characteristics The neoclassical theory argues that apart from regulations there are several additional industry characteristics such as technological changes and capacity utilization that are strongly associated with M&As (Gort, 1969; Mitchell and Mulherin, 1996; Andrade and Stafford, 2004; Jovanovic and Rousseau, 2002). Another influential view in the literature, known as the behavioural approach, argues that M&As are being driven by stock market conditions (Nelson, 1959; Stein, 1988, 1989, 1996; Baker and Wurgler, 2000; 2004; Jenter, 2005). Henceforth, various market related factors could potentially be considered, such as liquidity, profitability, growth, concentration, the level of economic development, 20

23 the size of the financial system and financial deepening. Some of these influences have been included as control variables in past studies. Market Liquidity Shleifer and Vishny (1992) show that in order for transactions to occur, buyers who intend to employ the assets in their first-best use must be relatively unconstrained. Schlingemann et al. (2002) reveal that industry-specific asset liquidity is important in determining which assets will be divested. Harford (2005) supports the neoclassical explanation that mergers occur in response to industry level economic, regulatory and technological shocks that require large-scale reallocation of assets, but suggests that shocks are not enough on their own, as capital liquidity is also required. Market Profitability The level or change in the profitability of the banking industry may also lead to higher acquisition activity as a result of attempts by banks to restructure or take advantage of investment opportunities that arise. Christensen and Montgomery (1981) show that firms in profitable industries tend to make more related acquisitions, while those from less profitable industries tend to be involved in unrelated acquisitions in order to improve their profit potential. Harford (2005) documents the existence of abnormally high changes in profitability prior to merger waves. However, Buch and Delong (2004a) find that the relative profitability of banking systems has little explanatory power for cross- 21

24 border merger activity, while Pasiouras and Zopounidis (2008) report a negative but insignificant effect of market profitability on bank M&As in Greece. Market Growth The growth of the market might also affect acquisition activity in two opposing ways. Firms might be attracted to engage in acquisitions within industries that have high growth rates, while in contrast low growth may indicate the need for restructuring in the industry, hence also leading to increased acquisition activity. Harford (2005) reports abnormally high growth measures (e.g. employees, sales) prior to waves, providing support to earlier studies claiming that firms are attracted to make acquisitions within industries with high growth rates (Christensen and Montgomery, 1981; Schoenberg and Reeves, 1999). However, evidence from the US and Greek banking sectors suggest that market growth is not a significant determinant of acquisition likelihood (Hannan and Rhoades, 1987; Pasiouras and Zopounidis, 2008). Kohler (2009) includes population as an indicator of market potential but finds its effect mostly insignificant on the probability of being acquired, specifically on domestic acquisitions within EU-25. Concentration Partly associated with the influence of regulations is market concentration, as anti-trust authorities try to prevent M&As if increases in concentration are expected to result in excessive increases in market power. According to Hannan and Piloff (2009), the likely impact of market concentration on acquisition likelihood depends on the degree to which the acquirer can exploit market power more efficiently than the target. They find the 22

25 effect of market concentration to be insignificant in most of the cases, while Hannan and Rhoades (1987) find the effect negative and significant in explaining the likelihood of inmarket acquisitions. Moore (1996) finds a positive and significant relationship between the probability of acquisition and market concentration for out-market acquisitions, but not for in-market acquisitions. Kohler (2009) finds the effect of market concentration positive and highly significant for both domestic and cross-border targets. On the other hand, Wheelock and Wilson (2004) and Pasiouras and Zopounidis (2008) report a negative relationship, while the results of Pasiouras and Gaganis (2009) are mixed. However, Hernando et al (2009) find that the effect of market concentration negative for domestic targets and positive for cross-border targets, although the overall effect in their combined sample is insignificant. They conclude that while domestic takeovers are less likely in more concentrated markets, the possibility of high rents and weak anti-trust concerns of national authorities with regard to foreign acquirers make is more likely that concentration has a positive effect on the probability of being acquired in cross-border takeovers. Economic Development The investment decision of banks can also be influenced by the overall economic conditions in which they operate. At one hand, banks could be involved in M&As during periods of boom to enhance their power and take advantage of the profit opportunities that arise. On the other hand, banks could be involved in M&As during periods of recession to be restructured and to avoid financial distress. Alternatively, bank may expand into regions with lower per-capita income because of higher income potential or 23

26 economic convergence expected from further integration (Lanine and Vander Vennett, 2007). Rossi and Volpin (2004) find the level of per capita GNP to be significant and positively related to the volume of M&As, but GDP growth to be negatively related and significant in four of their six specifications. Buch and DeLong (2004a) find the effect of GDP per capita of the acquirer country significantly positive, while that of the target country insignificant. Buch and DeLong (2004b) report that GDP has a positive and significant impact on the number of international bank mergers in either target or acquirers country. Lenine and Vander Vennet (2007) report an insignificant influence of per capita GDP growth (as with other macroeconomic influences) while Kohler (2009) finds the effect of target country s per capita GDP mostly negative and significant on the probability of being acquired. Size of the Banking Industry The size of the banking industry might have an impact on banks interest margins and profits (Demirguc-Kunt and Huizinga, 1999, 2000), their opportunities to achieve economies of scale (Buch and DeLong, 2004a) and consequently their M&As decisions. Diaz et al. (2004) examine the change in profitability of EU banks that were involved in acquisitions and report that the size of the banking sector has a negative and significant impact on profitability. Buch and DeLong (2004a) find that the size of the target country s banking systems has a negative impact on the probability of the merger, suggesting that banks do not invest in markets that have established a relatively large banking sector. Furthermore, De Nicolo (2000) argues that insolvency risk is lower in more developed financial markets, because these markets provide more financial 24

27 instruments that are more liquid than in developing markets. This leads us to the hypothesis that banks could use these financial instruments for diversification purposes rather than being involved in M&As. On the other hand, Demirguc-Kunt and Huizinga (1999, 2000) suggest that the lower interest margins in larger banking sectors might be related to increased competition. Hence, banks in these sectors might see M&As as a way to increase their power and competitiveness. Financial Deepening As previously mentioned, the behavioural approach states that stock market conditions might affect M&As because of valuation waves. Central to this hypothesis is the argument that bull markets encourage managers of firms with overvalued stock to use their stock to acquire undervalued targets. Verter (2002) and Giovanni (2005) among others confirm that stock market valuations and the size of the stock market are correlated with merger activity. Rhodes-Kropf and Viswanathan (2004) argue that the naïve explanation that overvalued bidders wish to use stock is incomplete because targets should not be eager to accept stock. They show that potential market value deviations from fundamental values on both sides of the transaction can rationally lead to a correlation between stock merger activity and market valuation. Shleifer and Vishny (2003) propose an alternative theory of acquisitions, which in a sense is the opposite of Roll s (1986) hubris hypothesis, by arguing that managers rationally respond to less than rational markets. Specifically they argue that since financial markets are inefficient, so some firms are valued incorrectly. Managers, on the other hand, who are completely rational, can understand stock market inefficiencies, and therefore take advantage through 25

28 M&As. Rhodes-Kropf et al. (2005) support the idea that misvaluation drives mergers. However, in contrast to these studies, Rossi and Volpin (2004) who examine domestic, foreign and hostile deals do not find any evidence to support that stock market return has an impact on M&As. III. RESEARCH DESIGN Data The sample used in this study consists of annual observations on the universe of commercial banks operating in the EU-15 with available financial data in Bankscope between 1996 and This gave us a sample of 1,407 banks with 5,986 observations, distinguished according to status (i.e. target, acquirer, non-involved), country and year as shown in Table I. 11 The geographical coverage of banks is as follows: Austria (63), Belgium (52), Denmark (70), Finland (9), France (278), Germany (226), Greece (24), Ireland (35), Italy (159), Luxembourg (132), Netherlands (51), Portugal (34), Spain (106), Sweden (16), and UK (152). [Insert Table I Around Here] 10 Only commercial banks were considered to avoid comparison problems among different type of banks (e.g. cooperative, savings, etc). Furthermore, as Demirguc-Kunt et al (2004) point out, since the WB regulatory data are for commercial banks, it is more appropriate to focus on commercial banks. 11 Targets and acquirers were identified in Bankscope, Bankersalmanac and Zephyr databases. Considering that acquisitions can take time to complete, we assume, as in previous studies (e.g. Hannan and Rhoades, 1987; Wheelock and Wilson, 2004) that acquisitions completed during year t, reflect their characteristics during year t-1. Thus for an acquisition that occurred in 1997 we use observations on variables from 1996, and so on. The sample is unbalanced in the sense that a complete panel of data is not available for each bank in the sample. Hence, the total number of banks with observed data in each year is lower than 1,

29 Many previous studies have followed a matched paired technique (e.g. Powell, 1997; Ali-Yrkko et al., 2005), where a sample of non-acquired firms is usually drawn by matching against the sample of acquired firms on the basis of company size, industry sector, and/or year of acquisition. While the advantage of this sampling procedure is that it helps to reduce the cost of data collection, a matched sample is limited in permitting investigation of the effects of industry sector differences. We have therefore chosen an unmatched sample that allows us to appropriately condition for differences in regulatory and environmental factors across countries. 12 Variables Table II summarises the set of 21 independent variables we use in the regressions below, classified for the purpose of discussion below as bank specific, regulatory and market related variables. [Insert Table II Around Here] Bank specific variables Seven bank specific ratios are chosen to represent the dimensions of capital strength, profitability, expenses management, loan activity, liquidity, size and growth. Capital strength is measured by the equity to total assets (EQAS) ratio, while profitability is measured with return on average equity (ROE). Efficiency in expenses management is represented by the cost to income ratio (COST), with higher values indicating less 12 Studies that use an unmatched sample include Hannan and Rhoades (1987), Lennox (1999), Jayaraman et al. (2002), Worthington (2004), Pasiouras and Zopounidis (2008) among others. 27